Why digitalization projects fail (and it is rarely about the technology)

Most digitalization projects do not fail on technology.

They fail because no one in the organization was allowed to care about the whole.

I have seen it in three different worlds. In a startup with ten employees. In public sector with a thousand. In a large corporation with a hundred thousand. The mechanism is the same every time.

Procurement buys the cheapest option. Those are the guidelines. The operations manager picks the easiest to implement. That is realistic. The IT manager picks the most secure. That is reasonable. The end user does whatever works in daily life. That is rational. Everyone does the right thing. The system does not work.

There is an argument I call the normal distribution argument. In a startup with ten people you can assemble ten brilliant minds. In an organization with ten thousand employees, the collective intelligence converges toward the mean. That is not an insult. It is statistics.

It means that decisions in large organizations are rarely very good or very bad. They are mediocre. The system is not chosen because it is right. It is chosen because it is not wrong. No single person has the mandate, the information or the incentive to optimize the whole. Everyone optimizes their part.

This is suboptimization. And it is not an exception. It is how large organizations work.

Now add AI on top of that.

The most common narrative right now: AI will make processes more efficient, remove friction and provide better decision support. Maybe that is true in theory. In practice, AI accelerates exactly the same pattern.

If procurement already buys the cheapest, AI does it faster. If the end user already ignores the system, AI gives them a fancier system to ignore. If nobody asked the frontline staff before, AI makes it easier not to. AI does not fix suboptimization. AI accelerates it.

That does not mean AI is pointless. It means that technology was never the problem. The problem is that no one owns the whole. The problem is that the incentive structure rewards partial optimization. The problem is that the question was wrong before the project even started.

I have seen projects that succeeded. What set them apart was not better technology, bigger budgets or more capable project managers. It was that someone had asked the right question before choosing tools. And that person had talked to the people who actually do the work.

That is still the cheapest step in the entire process. And it is still the step most often skipped.